Breakout 2B-3

Report
Actuarial Software Implementation
From Planning to Success
Actuarial Society of Greater New York
November 18, 2013
Trevor Howes, FCIA, FSA, MAAA
VP & Actuary, GGY AXIS
1
Agenda
• Introduction
• Scope of project
• Systems implementation planning
• IT considerations
• A useful resource: PBA Implementation Guide
Challenges Impacting Actuarial Software
Rapidly changing & new risks
Multi-core
chips
Rapid
changes
Technology
Innovation
Economic
.
Enviroment
Scarce capital; risk aversion
Volatile markets
IT
support
Low interest rates
HR costs
PBA
Actuarial
Models
Actuarial
support
Regulators
C3P2,3
MAR
ORSA
External
Stakeholders
Insurance Industry
New products
and pricing
Ratings
agencies
???
Demand for
higher
dividends
or ROC
SEC/FASB
SOX
???
IASB/IAIS
convergence
FASB 133, 157, SOP 03-1, etc.
3
Handle conflicting priorities in model design
Increase model realism
Increase model flexibility
•
•
•
•
• Multiple applications
• Production/ad hoc use
• Constant evolution as
needs, technology change
More detail, granularity
Asset/liability interaction
Holistic, integrated models
Reduce/control shortcuts
Reduce model expense
• Reduce maintenance and
validation effort
• Use technology efficiently
• Optional model efficiency
techniques
Improve model control
• Reduce model risk
• Assure run-time reliability
• Support auditability and
transparency standards
4
The Push for a Single Multi-function System
Pricing & Product
Development
Capital, Earnings &
Experience Analysis
Valuation, CFT &
Compliance
Projections, ALM &
Business Planning
5
Types of Actuarial System Implementations
Many possible variations…
• New system vs. conversion/upgrade to existing system
• Narrow Scope (single product & purpose) vs. Large Scope
(multiple purpose, multiple lines of business)
• Source of system: user vs. vendor written
• Approach to system maintenance: open vs. closed code ?
o Fully open
User has access to and
maintains all source code
o Partially open
Portions of code locked;
some user programming
o Fully closed
Only vendor maintains
source code
6
Why are Systems Implementations Challenging?
• Increasing complexity and precision of models
• Desire to improve capabilities, provide more analysis,
improve consistency, move to convergent systems
• Large diverse inforce product portfolios issued over past
decades and by various entities
• Need to develop, test, implement new system while
continuing to maintain existing systems and processes
• Systems implementations are major, resource intensive
projects with potentially significant risks …
but may be essential to transform the actuarial function and
to cope with emerging demands more efficiently
7
What are the Key Success Factors?
1. Expert knowledge of the company’s
systems, data and processes
2. Expert knowledge of the company’s
products
3. Expert knowledge of new software
including best practices for model
development
4. Appropriate planning and project
management
5. Appropriate time and resources to
work on the project
6. Commitment from senior levels of
the organization
1. Create a
realistic vision
of the
intended final
solution
2. Effectively
manage the
transition
process
8
Actuarial Models vs. Enterprise Applications
(Actuarial modeling engine)
Actuarial Models vs. Enterprise Applications
(Enterprise Systems and Databases)
Source
Stored Results
(Actuarial modeling engine)
Actuarial Models vs. Enterprise Applications
(Enterprise Systems and Databases)
Source
Stored Results
(Actuarial modeling engine)
(IT Infrastructure)
Software: Operating system / Network software / Grid Manager / Database Managers
Hardware: Desktop/Laptop/Servers/Storage Devices
The System Solution is more than the System
PBA Solutions will be much more than software
“The challenges in a PBA world will be solved through
collaboration and people exercising wisdom and judgment to
make actuarial and business decisions in the face of
ambiguity.” - SOA PBA Implementation Guide, pg 120.
13
Create Vision of System Solution
• Vision of future system solution is not a stand-alone
application, but part of enterprise level system
• Need to understand and plan for all related elements not just
actuarial modeling functionality
• Data input; data output; validation; audit; governance
• System solution includes the people; consider not just
actuarial roles of model maintenance/model use, but others
roles such as management and IT support
• Need to plan for staff transition and training during project
• Documentation must be created and maintained as model is
developed, not added after
Planning in a New and Uncertain Environment
• Systems implementations involve changes in scope,
capabilities, process; not just change from one software
product/solution to another
• Different software systems will require different approaches
to system structure, processing steps, data flow, report
generation, auditing support, etc.
• Not a drop in replacement of previous software
• Need to learn new software’s capabilities, strengths first
• Then look for opportunities to improve processes, and
address broader objectives and goals
15
Retain Flexibility in Model and Plan
• Expect ongoing changes in environment, vision and plan
• as project continues and
• after completion
• Future models will need to evolve and adapt as
regulatory and accounting standards, management
needs continue to change change
• Build a flexible model framework that can be
adapted after implementation phase completed
• Careful consideration of model governance issues
• Maybe separate initial project to design governance policy?
• Expect to revisit the plan itself as it progresses
16
Planning in a New and Uncertain Environment
• Initial planning should be
• Kept at reasonably high level
• Intended for preliminary assessment of resources and
organization, identification of potential risks and bottlenecks
• Plan project in phases, including pilot study (prototype) that
implements and tests full end-to-end process;
• Need for changes in model structure may become evident and
more easily implemented while model still small
• Communicate realistic expectations up front and repeatedly
through project
• Also plan for inevitable changes in the project team
• Training may be a continuous requirement
• Documentation as work continues is critical
17
Involve IT in Project Planning
• Need to involve IT in planning as early as possible
• New modeling demands require more processing power for
ongoing testing and implementation of assumption changes as
well as for production runs
• IT is often viewed as an adversary instead of a teammate
• Planning, sourcing, installing appropriate hardware takes time
• Actuarial models are not similar to typical insurance systems
• New software may be quite different than previous system
• Involve and trust the vendor for advice on specific hardware
specifications that will provide value for dollars spent
• Consider option to use private or public cloud instead of
company hosting either temporarily or permanently
• IT may assist in model governance planning
18
PBA Implementation Guide - Roadmaps
A Road Map conveys planning information identifying “what,” “how much,”
“who,” “when,” and “how.” A Road Map is:
• A multiyear integrated set of projects or initiatives and prioritizations
• An articulation of the right steps and resources
• A top-down high-level estimate of projected work efforts, duration
timelines, and all-in costs
• A prioritization framework to evaluate alternatives and their cost-benefits
and to translate implementation requirements into tangible steps
• A platform to provide focus and to facilitate ongoing analysis of priorities,
timelines, and investments
A Road Map’s initial purpose may be in securing management’s approval
and a PBA budget.
(Extracted from PBA Implementation Guide, pg. 10)
19
PBA Implementation Guide - Roadmaps
The Road Map Guide … has the following sections:
Introduction
1 Assumption Setting
2 Inputs
3 Model Platforms (Throughputs)
4 Outputs
5 Technology and Systems
6 Actuarial Organization
7 Potential Road Map Initiatives
(PBA Implementation Guide, pg. 96)
20
PBA Implementation Guide – Case Study Roadmaps
• 6 illustrative Case Study Road Maps (templates)
• Implicit considerations included:
•
•
•
•
•
•
•
•
•
•
recognition of priorities
ease of implementation
current framework and gaps
importance
costs
realized benefits
work effort
resource availability
skill set gaps
conflicts
• bottlenecks
• dependencies
• discontinuous work efforts
and business or project
continuities
• execution risks
• coordination with other
company projects
• balancing PBA work efforts
with day-to-day activities.
(PBA Implementation Guide, pg. 26)
21
Sample Case Study Roadmap
22
Sample Road Map Initiatives
23
Reference:
• PBA Implementation Guide - a 2013 Research report
sponsored by Society of Actuaries, prepared by:
Tim Cardinal, FSA, CERA, MAAA
Steve Stockman, ASA, MAAA
• For the full report, go to:
http://www.soa.org/Files/Research/Projects/research-2013-pbaimplementation-guide.pdf
24

similar documents